www.nature.com/scientificreports
OPEN
received: 22 August 2016 accepted: 04 January 2017 Published: 07 February 2017
Electromagnetic pulsed thermography for natural cracks inspection Yunlai Gao1,2, Gui Yun Tian1,3, Ping Wang2, Haitao Wang2, Bin Gao3, Wai Lok Woo1 & Kongjing Li1 Emerging integrated sensing and monitoring of material degradation and cracks are increasingly required for characterizing the structural integrity and safety of infrastructure. However, most conventional nondestructive evaluation (NDE) methods are based on single modality sensing which is not adequate to evaluate structural integrity and natural cracks. This paper proposed electromagnetic pulsed thermography for fast and comprehensive defect characterization. It hybrids multiple physical phenomena i.e. magnetic flux leakage, induced eddy current and induction heating linking to physics as well as signal processing algorithms to provide abundant information of material properties and defects. New features are proposed using 1st derivation that reflects multiphysics spatial and temporal behaviors to enhance the detection of cracks with different orientations. Promising results that robust to lift-off changes and invariant features for artificial and natural cracks detection have been demonstrated that the proposed method significantly improves defect detectability. It opens up multiphysics sensing and integrated NDE with potential impact for natural understanding and better quantitative evaluation of natural cracks including stress corrosion crack (SCC) and rolling contact fatigue (RCF). Emerging integrated techniques, sensing and monitoring materials degradation are increasingly required for characterizing the structural integrity and safety of infrastructure1,2. Nondestructive evaluation (NDE)3,4 is a typical effective sensing approach to recognize material characteristics and structural degradation without destroying the serviceability of a component or system. However, conventional NDE based on single modality sensing is not adequate to evaluate structural integrity with the required spatial resolution, coverage and accuracy4–8. A single NDE method using one modality of physics such as ultrasonic, electrical, magnetic, optical, thermal and radiography will not always be sufficient for the purpose of complete defects evaluation. To overcome the above problems, the use of multiple modality sensing and fusion5,8–10 in a complementary manner paths the way to enable a comprehensive understanding of material and structural characteristics. Integrating sensing techniques such as electromagnetic acoustic transducer (EMAT)8, infrared and optic thermography9, and magneto-optical visualization10 provide a promising multiphysics problem-solving approach to enhance the NDE performance for material and defect evaluation. The sensing techniques based on the electromagnetic, i.e. eddy current7, magnetic flux leakage (MFL)6, and alternating current field measurement (ACFM)11 have been widely applied in NDE. Eddy current pulsed thermography (ECPT)12–14 is an emerging multiple modality NDE technique for conductive material which combines both advantages of pulsed eddy current (PEC)7 and infrared thermography15. The material properties such as electrical conductivity, magnetic permeability and thermal conductivity are used to identify and evaluate the features of interest. Material characteristics and structural stress, fatigue or damages are efficiently recorded and demonstrated through thermal image sequences can be evaluated by Joule heating via eddy current, heat conduction and infrared thermography12,13. Thermal patterns including the contrast against background of the tested object are analyzed by image processing and feature extraction e.g. differential absolute contrast (DAC)16, thermographic signal reconstruction (TSR)12, principle component thermography(PCT)17 and pulsed phase thermography (PPT)18 methods for defect characterization12,14,19. The superior performances of ECPT such 1
School of Electrical and Electronic Engineering, Newcastle University, Newcastle upon Tyne, NE1 7RU, UK. 2College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China. 3School of Automation Engineering, University of Electronics Science and Technology of China, Chengdu 611731, China. Correspondence and requests for materials should be addressed to G.Y.T. (email: [email protected]) Scientific Reports | 7:42073 | DOI: 10.1038/srep42073
1
www.nature.com/scientificreports/
Figure 1. Fabrication of electromagnetic pulsed thermography system. In (a), the schematic diagram with well-established excitation and sensing configuration is depicted. The U-shaped ferrite-core act as inductor; the distance between the pole shoes and the ferromagnetic test object is defined as yoke lift-off. The size of the ferrite-core and the sample are also illustrated for further multiphysics modeling, simulation and experimental arrangement. In (b), the experimental system is presented to show further capabilities of the approach for material properties and defect characterization (Details of operation are provided below).
as non-contact, high resolution, fast detection in a large area with rich transient information enable extensive research and NDE applications in recent years14,15. Wilson et al.15 proposed the PEC thermography to detect multiple cracks of rail rolling contact fatigue (RCF). Cheng20 and He21 ‘lighted’ and detected impact damages of composite material using ECPT. He et al.22 investigated the ECPT for detection of corrosion blister in mild steel. Li et al.23 used the ECPT for bond wire state detection in electrical modules. Tian et al.24 applied the ECPT for early fatigue evaluation of gear. Yin et al.13 reported the physical interpretation of ECPT and discussed the links between mathematical and physical models. Gao et al.14 extracted spatial and time patterns for automatic NDE based on the ECPT transient thermal sequences. However, the problems of in-homogenous heating, limited heating area and blocking effect of coil24–26 in reflection mode ECPT are still challenging the accurate material and defect quantitative characterization. Lahiri et al.27 reported low frequency alternating magnetic field for thermographic NDE of defect with large area. Jäckel et al.28 proposed an electromagnet yoke for external magnetic field to enhance the crack detection contrast by induction thermography. Hansen et al.29 designed an asymmetric induction coil for generating uniform heating. Netzelmann et al.30 developed a measurement system with induction generator scanning for rail surface defects inspection at different speed up to 15 km/h. Shepard et al.31 described advances and the analysis of pulsed thermographic data for defect detection with increased spatial and temporal resolution. The implementation of uniform heating in a large area together with a wide open-view region for defect imaging is required to overcome the above problems of ECPT. To enhance the NDE performance of the ECPT, more physics integration including but not limited to the eddy current with thermal imaging are urgently required for comprehensive NDE of natural multiple defects, larger area detection of structures with free-form surface and evaluations of carbon fiber reinforced plastics (CFRP) flaws20,21 and RCFs25 etc. In this work, we developed an electromagnetic pulsed thermography using two physical phenomena of magnetic field and eddy current for induced heating for fast and comprehensive defect characterization. The response signals are interpreted to provide abundant material and defective information through temporal and spatial responses. The proposed method with multiphysics sensing and interpretation are based on the integration responses of different NDE techniques, e.g. MFL or EC, and a fusion of different physical phenomena for heating, these includes (1) induced eddy current generates Joule heating, (2) alternating magnetization/demagnetization produce hysteresis loss for heating, and (3) leakage magnetic flux with stray loss for different time and amplitude responses32. The thermal patterns captured by infrared imaging enables the visualization of eddy current, magnetic field behaviors and their heating effect. In addition, the new feature of 1st derivation thermal pattern calculation is proposed for enhancing natural cracks inspection. The system operational principle is examined in simulation with interactions of multiple physical field behavior. New features are extracted and confirmed in experiments to estimate crack orientations through the response time and speed, which reflect different percentage contributions of magnetic flux leakage and eddy current for heating as well as defect evaluation. Different spatial and temporal responses of multiphysics reveal the promising relative uniform field excitation together with an open-view imaging for accurate defect detection and characterization. The advanced performance of the proposed method e.g. robust to lift-off changes and effective for defect orientation and depth estimation opens up multiphysics sensing and integrated NDE with potential significant impact for fast quantitative evaluation.
Methods
Implementation of electromagnetic pulsed thermography system. To implement the multiphysics sensing approach for comprehensive NDE, a new electromagnetic pulsed thermography system is illustrated with a schematic diagram in Fig. 1a and b. The proposed system is based on the comparison studies of the magnetic flux leakage and ECPT techniques and combination of both advantages. A high power and high frequency alternating electrical current is generated by a pulse generator and induction heater to drive the excitation coil. A very intense and rapidly changing magnetic flux is produced in the space within the coil. Compared with the previous ECPT13,15, the most significant improvement of this procedure is using a ferrite-core to concentrate most of the Scientific Reports | 7:42073 | DOI: 10.1038/srep42073
2
www.nature.com/scientificreports/ magnetic flux into a magnetic circuit and guiding them to flow into the test object with a broad and uniform field distribution. In addition, the use of a ferrite-core (magnet yoke) enables the decrease of magnetic resistance in the magnetic circuit and enhances the magnetic flux intensity in the ferromagnetic test object for efficient induction heating. Moreover, the local region between two pole shoes of the magnet yoke provides a wide open-view area for defect detection and visualization by full coverage infrared imaging, which will significantly benefit for quantitative NDE. Thermal image sequences including multiphysics spatial and temporal responses and defect characteristics are transmitted to a computer for signal processing and defect features extraction. The characteristics of the proposed electromagnetic pulsed thermography system enable the integration of multiple physical phenomena for material and defect characterization. These includes: (i) Magnetic flux leakage with stray loss: Any geometrical discontinuity or local anomalies existing in the measured area lead to the magnetic flux leakage to air around the defective area due to the abrupt changes of magnetic permeability, which is the MFL principle6,33. Around the defective area, the local magnetic field distribution inside of the test object will be disrupted and represented as high/low magnetic flux density. Based on the stray loss due to the leakage of magnetic flux, this paper uses remaining magnetic field inside of test object around the defective area to induce different eddy current and inductive heating responses for defect detection. (ii) Induced eddy current: significant amount of eddy currents will be induced by rapidly alternating magnetic field and orthogonally distributed against the magnetic field lines in the surface and subsurface region of test object13,24. Similar to the MFL, the features of interest manifested as an abrupt change of electrical conductivity will lead to the disturbance of eddy current field and represent as increased/decreased density around the defective area. Based on the eddy current loss, the heating effect of eddy current is utilized for the visualization of the eddy current field distribution and intensity behaviors for the defect detection. (iii) Induction heating with multiphysics34,35: The induced eddy current will give rise to local resistive heat by Joule heating13. Due to the hysteresis effect36,37 of the ferromagnetic material, it naturally offers resistance to intense alternating magnetic field because of the repeatedly magnetize/demagnetize processes. The rapid periodic creation and annihilation of magnetic domains by domain wall movements38,39 cause considerable internal friction damping, magnetostrictively moved dislocation and heating inside the material, which is potential for speedy and efficient thermographic NDE due to the secondary source of heat40,41. Additionally, the leakage of magnetic flux with stray loss also induce different inductive heating response due to magnetic field distribution and intensity variation inside of test object. Above all, the magnetic flux and eddy current field diversion and intensity changes around the defective area will lead to the different heat density and thermal contrast between defective and defect-free area, which makes defect visible using an IR camera. The physical field distribution and dynamic behaviors of magnetic field and eddy current field are lighted by the above heating effect and visualized through thermal imaging. The dynamic thermal behavior also can be used to characterize defective features because of the heat conduction and material thermal conductivity variation13,24. When the proposed system is installed on an inspection car which is rapidly driving on the rail, the permanently electromagnetic pulsed heating and thermal imaging are continually working for natural cracks location and characterization.
Multiphysics modelling and simulations. In this section, the interactions behavior of the multiple physical field by using multiphysics modeling and simulations conducted with COMSOL Multiphysics 4.3b6,25 is conducted. A three-dimensional (3D) model is built on account of the proposed system configuration in Fig. 1a. The radius and wire diameter of the excitation coil are 25 mm and 6.35 mm, respectively. The dimensions of the U-shaped ferrite-core are 120 × 30 × 90 mm3 in length, width and height with a vacant region of 60 × 30 × 60 mm3 between two pole shoes of the ferrite-core. The steel sample with 200 × 150 × 10 mm3 includes three oriented slots with the orientations 0 degree, 45 degree, 90 degree and dimensions of 10 × 1 × 5 mm3 in length, width and depth, respectively. The distance between the pole shoes of the U-shaped ferrite-core and the ferromagnetic test object is 1 mm, which is defined as yoke lift-off. The materials properties of the ferrite-core, steel and air are setup based on their real parameters such as relative permeability 5000, 800, 1; electrical conductivity (unit S/m) 0.01, 4.03 × 106, 0.01 and thermal conductivity (unit W/(m × k)) 2.90, 44.50, 2.57 × 10−2. The fine mesh was generated within the problem regions of interest, e.g. three slots areas. The mesh quality was improved in the steel sample to achieve accurate multiple physical fields distribution without too much sacrifice of computing time. The entire 3D model was divided into 215571 tetrahedral elements for the finite element calculation. The calculation is implemented based on the AC/DC module of COMSOL with the maximum number of iterations 10. The simulation of 3D model took about 28 minutes in a typical 2.2 GHz Intel Core i5 processor computer with 8 G memory. Electrical current 380 A with frequency 256 kHz is driven into the excitation coil to generate intense alternating magnetic flux in the space within the coil. Most of the magnetic flux are concentrated in the ferrite-core and guided into the test object for inducing eddy current and heating in order to identify defective characteristics. Duration of induction heating is setup with 300 ms, which is enough to generate effective thermal response and contrast for material properties and defect characterization. Experimental set-up and samples. The experimental system of the proposed electromagnetic pulsed thermography is developed as shown in Fig. 1b. It consists of a function generator, heating source devices, excitation coil, ferrite-core, test sample and an IR camera as well as a PC for signal processing. A function generator Agilent 33500B is used to produce a pulsed signal to trigger and control the operation of the IR camera and heating devices. A precision induction heating device, Easyheat 224 from Cheltenham Induction Heating, Ltd., together with a cooler and a workhead, is used for coil excitation, with a maximum excitation power of 2.4 kW, a maximum current of 400 Arms, and an excitation frequency range of 150–400 kHz (380 Arms and 256 kHz are used in this study). The excitation coil is made of 6.35 mm diameter high-conductivity hollow copper tube where water cooling is implemented to counteract direct heating of the coil. To fit the U-shaped ferrite-core structure, the coil
Scientific Reports | 7:42073 | DOI: 10.1038/srep42073
3
www.nature.com/scientificreports/
Figure 2. Simulation results of multiple physical field interactions behavior. In (a), the stimulated magnetic flux vectors (green) and induced eddy current vectors (red) in the measured area are depicted together with induction heat distribution. In (b), three physical field interactions behavior to transversal, longitudinal and oblique oriented cracks are illustrated to characterize the defect. is designed as a rectangular shape and wound on the magnet yoke. The ferrite-core is made of MnZn ferrite material with parameters and sizes identical to simulation. A mild steel sample13 containing a narrow, surface breaking slot in length 30 mm, width 0.5 mm and depth 6 mm is employed in the experimental tests. Through changing the relative position between ferrite-core and sample, various crack orientations to the excitation are detected and visualized by an IR camera. A short piece of rail track sample including rolling contact fatigue with natural multiple cracks is also employed for test. For defect depth estimation, three pieces of steel sample with 5 mm thickness, 150 mm length and 50 mm width are employed for test. Three slots of which the depths are 2 mm, 3 mm and 4 mm, respectively, with same defect length 10 mm and width 1 mm. The state-of-the-art IR system Flir SC7500 is used to record the temperature change and thermodynamics behavior, which is a Stirling cooled camera with a 320 × 256 array of 1.5–5 µm InSb detector. It has a sensitivity of
OPEN
received: 22 August 2016 accepted: 04 January 2017 Published: 07 February 2017
Electromagnetic pulsed thermography for natural cracks inspection Yunlai Gao1,2, Gui Yun Tian1,3, Ping Wang2, Haitao Wang2, Bin Gao3, Wai Lok Woo1 & Kongjing Li1 Emerging integrated sensing and monitoring of material degradation and cracks are increasingly required for characterizing the structural integrity and safety of infrastructure. However, most conventional nondestructive evaluation (NDE) methods are based on single modality sensing which is not adequate to evaluate structural integrity and natural cracks. This paper proposed electromagnetic pulsed thermography for fast and comprehensive defect characterization. It hybrids multiple physical phenomena i.e. magnetic flux leakage, induced eddy current and induction heating linking to physics as well as signal processing algorithms to provide abundant information of material properties and defects. New features are proposed using 1st derivation that reflects multiphysics spatial and temporal behaviors to enhance the detection of cracks with different orientations. Promising results that robust to lift-off changes and invariant features for artificial and natural cracks detection have been demonstrated that the proposed method significantly improves defect detectability. It opens up multiphysics sensing and integrated NDE with potential impact for natural understanding and better quantitative evaluation of natural cracks including stress corrosion crack (SCC) and rolling contact fatigue (RCF). Emerging integrated techniques, sensing and monitoring materials degradation are increasingly required for characterizing the structural integrity and safety of infrastructure1,2. Nondestructive evaluation (NDE)3,4 is a typical effective sensing approach to recognize material characteristics and structural degradation without destroying the serviceability of a component or system. However, conventional NDE based on single modality sensing is not adequate to evaluate structural integrity with the required spatial resolution, coverage and accuracy4–8. A single NDE method using one modality of physics such as ultrasonic, electrical, magnetic, optical, thermal and radiography will not always be sufficient for the purpose of complete defects evaluation. To overcome the above problems, the use of multiple modality sensing and fusion5,8–10 in a complementary manner paths the way to enable a comprehensive understanding of material and structural characteristics. Integrating sensing techniques such as electromagnetic acoustic transducer (EMAT)8, infrared and optic thermography9, and magneto-optical visualization10 provide a promising multiphysics problem-solving approach to enhance the NDE performance for material and defect evaluation. The sensing techniques based on the electromagnetic, i.e. eddy current7, magnetic flux leakage (MFL)6, and alternating current field measurement (ACFM)11 have been widely applied in NDE. Eddy current pulsed thermography (ECPT)12–14 is an emerging multiple modality NDE technique for conductive material which combines both advantages of pulsed eddy current (PEC)7 and infrared thermography15. The material properties such as electrical conductivity, magnetic permeability and thermal conductivity are used to identify and evaluate the features of interest. Material characteristics and structural stress, fatigue or damages are efficiently recorded and demonstrated through thermal image sequences can be evaluated by Joule heating via eddy current, heat conduction and infrared thermography12,13. Thermal patterns including the contrast against background of the tested object are analyzed by image processing and feature extraction e.g. differential absolute contrast (DAC)16, thermographic signal reconstruction (TSR)12, principle component thermography(PCT)17 and pulsed phase thermography (PPT)18 methods for defect characterization12,14,19. The superior performances of ECPT such 1
School of Electrical and Electronic Engineering, Newcastle University, Newcastle upon Tyne, NE1 7RU, UK. 2College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China. 3School of Automation Engineering, University of Electronics Science and Technology of China, Chengdu 611731, China. Correspondence and requests for materials should be addressed to G.Y.T. (email: [email protected]) Scientific Reports | 7:42073 | DOI: 10.1038/srep42073
1
www.nature.com/scientificreports/
Figure 1. Fabrication of electromagnetic pulsed thermography system. In (a), the schematic diagram with well-established excitation and sensing configuration is depicted. The U-shaped ferrite-core act as inductor; the distance between the pole shoes and the ferromagnetic test object is defined as yoke lift-off. The size of the ferrite-core and the sample are also illustrated for further multiphysics modeling, simulation and experimental arrangement. In (b), the experimental system is presented to show further capabilities of the approach for material properties and defect characterization (Details of operation are provided below).
as non-contact, high resolution, fast detection in a large area with rich transient information enable extensive research and NDE applications in recent years14,15. Wilson et al.15 proposed the PEC thermography to detect multiple cracks of rail rolling contact fatigue (RCF). Cheng20 and He21 ‘lighted’ and detected impact damages of composite material using ECPT. He et al.22 investigated the ECPT for detection of corrosion blister in mild steel. Li et al.23 used the ECPT for bond wire state detection in electrical modules. Tian et al.24 applied the ECPT for early fatigue evaluation of gear. Yin et al.13 reported the physical interpretation of ECPT and discussed the links between mathematical and physical models. Gao et al.14 extracted spatial and time patterns for automatic NDE based on the ECPT transient thermal sequences. However, the problems of in-homogenous heating, limited heating area and blocking effect of coil24–26 in reflection mode ECPT are still challenging the accurate material and defect quantitative characterization. Lahiri et al.27 reported low frequency alternating magnetic field for thermographic NDE of defect with large area. Jäckel et al.28 proposed an electromagnet yoke for external magnetic field to enhance the crack detection contrast by induction thermography. Hansen et al.29 designed an asymmetric induction coil for generating uniform heating. Netzelmann et al.30 developed a measurement system with induction generator scanning for rail surface defects inspection at different speed up to 15 km/h. Shepard et al.31 described advances and the analysis of pulsed thermographic data for defect detection with increased spatial and temporal resolution. The implementation of uniform heating in a large area together with a wide open-view region for defect imaging is required to overcome the above problems of ECPT. To enhance the NDE performance of the ECPT, more physics integration including but not limited to the eddy current with thermal imaging are urgently required for comprehensive NDE of natural multiple defects, larger area detection of structures with free-form surface and evaluations of carbon fiber reinforced plastics (CFRP) flaws20,21 and RCFs25 etc. In this work, we developed an electromagnetic pulsed thermography using two physical phenomena of magnetic field and eddy current for induced heating for fast and comprehensive defect characterization. The response signals are interpreted to provide abundant material and defective information through temporal and spatial responses. The proposed method with multiphysics sensing and interpretation are based on the integration responses of different NDE techniques, e.g. MFL or EC, and a fusion of different physical phenomena for heating, these includes (1) induced eddy current generates Joule heating, (2) alternating magnetization/demagnetization produce hysteresis loss for heating, and (3) leakage magnetic flux with stray loss for different time and amplitude responses32. The thermal patterns captured by infrared imaging enables the visualization of eddy current, magnetic field behaviors and their heating effect. In addition, the new feature of 1st derivation thermal pattern calculation is proposed for enhancing natural cracks inspection. The system operational principle is examined in simulation with interactions of multiple physical field behavior. New features are extracted and confirmed in experiments to estimate crack orientations through the response time and speed, which reflect different percentage contributions of magnetic flux leakage and eddy current for heating as well as defect evaluation. Different spatial and temporal responses of multiphysics reveal the promising relative uniform field excitation together with an open-view imaging for accurate defect detection and characterization. The advanced performance of the proposed method e.g. robust to lift-off changes and effective for defect orientation and depth estimation opens up multiphysics sensing and integrated NDE with potential significant impact for fast quantitative evaluation.
Methods
Implementation of electromagnetic pulsed thermography system. To implement the multiphysics sensing approach for comprehensive NDE, a new electromagnetic pulsed thermography system is illustrated with a schematic diagram in Fig. 1a and b. The proposed system is based on the comparison studies of the magnetic flux leakage and ECPT techniques and combination of both advantages. A high power and high frequency alternating electrical current is generated by a pulse generator and induction heater to drive the excitation coil. A very intense and rapidly changing magnetic flux is produced in the space within the coil. Compared with the previous ECPT13,15, the most significant improvement of this procedure is using a ferrite-core to concentrate most of the Scientific Reports | 7:42073 | DOI: 10.1038/srep42073
2
www.nature.com/scientificreports/ magnetic flux into a magnetic circuit and guiding them to flow into the test object with a broad and uniform field distribution. In addition, the use of a ferrite-core (magnet yoke) enables the decrease of magnetic resistance in the magnetic circuit and enhances the magnetic flux intensity in the ferromagnetic test object for efficient induction heating. Moreover, the local region between two pole shoes of the magnet yoke provides a wide open-view area for defect detection and visualization by full coverage infrared imaging, which will significantly benefit for quantitative NDE. Thermal image sequences including multiphysics spatial and temporal responses and defect characteristics are transmitted to a computer for signal processing and defect features extraction. The characteristics of the proposed electromagnetic pulsed thermography system enable the integration of multiple physical phenomena for material and defect characterization. These includes: (i) Magnetic flux leakage with stray loss: Any geometrical discontinuity or local anomalies existing in the measured area lead to the magnetic flux leakage to air around the defective area due to the abrupt changes of magnetic permeability, which is the MFL principle6,33. Around the defective area, the local magnetic field distribution inside of the test object will be disrupted and represented as high/low magnetic flux density. Based on the stray loss due to the leakage of magnetic flux, this paper uses remaining magnetic field inside of test object around the defective area to induce different eddy current and inductive heating responses for defect detection. (ii) Induced eddy current: significant amount of eddy currents will be induced by rapidly alternating magnetic field and orthogonally distributed against the magnetic field lines in the surface and subsurface region of test object13,24. Similar to the MFL, the features of interest manifested as an abrupt change of electrical conductivity will lead to the disturbance of eddy current field and represent as increased/decreased density around the defective area. Based on the eddy current loss, the heating effect of eddy current is utilized for the visualization of the eddy current field distribution and intensity behaviors for the defect detection. (iii) Induction heating with multiphysics34,35: The induced eddy current will give rise to local resistive heat by Joule heating13. Due to the hysteresis effect36,37 of the ferromagnetic material, it naturally offers resistance to intense alternating magnetic field because of the repeatedly magnetize/demagnetize processes. The rapid periodic creation and annihilation of magnetic domains by domain wall movements38,39 cause considerable internal friction damping, magnetostrictively moved dislocation and heating inside the material, which is potential for speedy and efficient thermographic NDE due to the secondary source of heat40,41. Additionally, the leakage of magnetic flux with stray loss also induce different inductive heating response due to magnetic field distribution and intensity variation inside of test object. Above all, the magnetic flux and eddy current field diversion and intensity changes around the defective area will lead to the different heat density and thermal contrast between defective and defect-free area, which makes defect visible using an IR camera. The physical field distribution and dynamic behaviors of magnetic field and eddy current field are lighted by the above heating effect and visualized through thermal imaging. The dynamic thermal behavior also can be used to characterize defective features because of the heat conduction and material thermal conductivity variation13,24. When the proposed system is installed on an inspection car which is rapidly driving on the rail, the permanently electromagnetic pulsed heating and thermal imaging are continually working for natural cracks location and characterization.
Multiphysics modelling and simulations. In this section, the interactions behavior of the multiple physical field by using multiphysics modeling and simulations conducted with COMSOL Multiphysics 4.3b6,25 is conducted. A three-dimensional (3D) model is built on account of the proposed system configuration in Fig. 1a. The radius and wire diameter of the excitation coil are 25 mm and 6.35 mm, respectively. The dimensions of the U-shaped ferrite-core are 120 × 30 × 90 mm3 in length, width and height with a vacant region of 60 × 30 × 60 mm3 between two pole shoes of the ferrite-core. The steel sample with 200 × 150 × 10 mm3 includes three oriented slots with the orientations 0 degree, 45 degree, 90 degree and dimensions of 10 × 1 × 5 mm3 in length, width and depth, respectively. The distance between the pole shoes of the U-shaped ferrite-core and the ferromagnetic test object is 1 mm, which is defined as yoke lift-off. The materials properties of the ferrite-core, steel and air are setup based on their real parameters such as relative permeability 5000, 800, 1; electrical conductivity (unit S/m) 0.01, 4.03 × 106, 0.01 and thermal conductivity (unit W/(m × k)) 2.90, 44.50, 2.57 × 10−2. The fine mesh was generated within the problem regions of interest, e.g. three slots areas. The mesh quality was improved in the steel sample to achieve accurate multiple physical fields distribution without too much sacrifice of computing time. The entire 3D model was divided into 215571 tetrahedral elements for the finite element calculation. The calculation is implemented based on the AC/DC module of COMSOL with the maximum number of iterations 10. The simulation of 3D model took about 28 minutes in a typical 2.2 GHz Intel Core i5 processor computer with 8 G memory. Electrical current 380 A with frequency 256 kHz is driven into the excitation coil to generate intense alternating magnetic flux in the space within the coil. Most of the magnetic flux are concentrated in the ferrite-core and guided into the test object for inducing eddy current and heating in order to identify defective characteristics. Duration of induction heating is setup with 300 ms, which is enough to generate effective thermal response and contrast for material properties and defect characterization. Experimental set-up and samples. The experimental system of the proposed electromagnetic pulsed thermography is developed as shown in Fig. 1b. It consists of a function generator, heating source devices, excitation coil, ferrite-core, test sample and an IR camera as well as a PC for signal processing. A function generator Agilent 33500B is used to produce a pulsed signal to trigger and control the operation of the IR camera and heating devices. A precision induction heating device, Easyheat 224 from Cheltenham Induction Heating, Ltd., together with a cooler and a workhead, is used for coil excitation, with a maximum excitation power of 2.4 kW, a maximum current of 400 Arms, and an excitation frequency range of 150–400 kHz (380 Arms and 256 kHz are used in this study). The excitation coil is made of 6.35 mm diameter high-conductivity hollow copper tube where water cooling is implemented to counteract direct heating of the coil. To fit the U-shaped ferrite-core structure, the coil
Scientific Reports | 7:42073 | DOI: 10.1038/srep42073
3
www.nature.com/scientificreports/
Figure 2. Simulation results of multiple physical field interactions behavior. In (a), the stimulated magnetic flux vectors (green) and induced eddy current vectors (red) in the measured area are depicted together with induction heat distribution. In (b), three physical field interactions behavior to transversal, longitudinal and oblique oriented cracks are illustrated to characterize the defect. is designed as a rectangular shape and wound on the magnet yoke. The ferrite-core is made of MnZn ferrite material with parameters and sizes identical to simulation. A mild steel sample13 containing a narrow, surface breaking slot in length 30 mm, width 0.5 mm and depth 6 mm is employed in the experimental tests. Through changing the relative position between ferrite-core and sample, various crack orientations to the excitation are detected and visualized by an IR camera. A short piece of rail track sample including rolling contact fatigue with natural multiple cracks is also employed for test. For defect depth estimation, three pieces of steel sample with 5 mm thickness, 150 mm length and 50 mm width are employed for test. Three slots of which the depths are 2 mm, 3 mm and 4 mm, respectively, with same defect length 10 mm and width 1 mm. The state-of-the-art IR system Flir SC7500 is used to record the temperature change and thermodynamics behavior, which is a Stirling cooled camera with a 320 × 256 array of 1.5–5 µm InSb detector. It has a sensitivity of
